US2025294085A1PendingUtilityA1

Hardware assisted atomics support in datacenter fabrics

52
Assignee: ADVANCED MICRO DEVICES INCPriority: Mar 13, 2024Filed: Mar 13, 2024Published: Sep 18, 2025
Est. expiryMar 13, 2044(~17.7 yrs left)· nominal 20-yr term from priority
H04L 69/22H04L 41/14
52
PatentIndex Score
0
Cited by
0
References
0
Claims

Abstract

Embodiments herein describe generating packets that combine an atomic operation (e.g., an atomic fetch) with a data operation (e.g., a put). Previous remote atomics first transmit a packet to a remote node that provides a lock for the data. If the lock is granted, the node transmits another packet containing a data operation which can read or write data. However, the embodiments herein can use a relax range-based atomics where the packet uses a range to reserve space in a dataset (e.g., a buffer) at the destination node for the data operation.

Claims

exact text as granted — not AI-modified
What is claimed is: 
     
         1 . A computing system, comprising:
 one or more processors; and   memory storing an application which, when executed by the one or more processors, performs an operation, the operation comprising:
 generating a relaxed range for a data operation; 
 generating a packet that combines an atomic operation with the data operation; and 
 transmitting the packet on a network. 
   
     
     
         2 . The computing system of  claim 1 , wherein the relaxed range comprises a beginning address and an ending address related to a size of a payload of the data operation. 
     
     
         3 . The computing system of  claim 1 , wherein relaxed range comprises a tuple that includes a destination base address, a destination end address, and an ID of a destination node for the packet. 
     
     
         4 . The computing system of  claim 1 , wherein the atomic operation obtains a lock of data associated with the data operation. 
     
     
         5 . The computing system of  claim 4 , wherein the data operation is a put operation that stores data in a buffer at a destination node after the lock is obtained. 
     
     
         6 . The computing system of  claim 1 , wherein the packet contains an extended transport header (ETH) indicating that the packet contains an atomic operation. 
     
     
         7 . The computing system of  claim 6 , wherein the ETH comprises a header that includes an atomic virtual address that specifies an address for an atomic operation and a base virtual address that specifies a first memory word of a buffer where data should be read from, or written to, at a destination node. 
     
     
         8 . A method, comprising:
 generating a relaxed range for a data operation;   generating a packet that combines an atomic operation with the data operation; and   transmitting the packet on a network.   
     
     
         9 . The method of  claim 8 , wherein the relaxed range comprises a beginning address and an ending address related to a size of a payload of the data operation. 
     
     
         10 . The method of  claim 8 , wherein relaxed range comprises a tuple that includes a destination base address, a destination end address, and an ID of a destination node for the packet. 
     
     
         11 . The method of  claim 8 , wherein the atomic operation obtains a lock of data associated with the data operation. 
     
     
         12 . The method of  claim 11 , wherein the data operation is a put operation that stores data in a buffer at a destination node after the lock is obtained. 
     
     
         13 . The method of  claim 8 , wherein the packet contains an extended transport header (ETH) indicating that the packet contains an atomic operation. 
     
     
         14 . The method of  claim 13 , wherein the ETH comprises a header that includes an atomic virtual address that specifies an address for an atomic operation and a base virtual address that specifies a first memory word of a buffer where data should be read from, or written to, at a destination node. 
     
     
         15 . A non-transitory computer readable storage medium comprising computer readable program code embodied therewith, the computer readable program code executable by one or more computer processors to perform an operation, the operation comprising:
 generating a relaxed range for a data operation;   generating a packet that combines an atomic operation with the data operation; and   transmitting the packet on a network.   
     
     
         16 . The non-transitory computer readable storage medium of  claim 15 , wherein the relaxed range comprises a beginning address and an ending address related to a size of a payload of the data operation. 
     
     
         17 . The non-transitory computer readable storage medium of  claim 15 , wherein relaxed range comprises a tuple that includes a destination base address, a destination end address, and an ID of a destination node for the packet. 
     
     
         18 . The non-transitory computer readable storage medium of  claim 15 , wherein the atomic operation obtains a lock of data associated with the data operation, wherein the data operation is a put operation that stores data in a buffer at a destination node after the lock is obtained. 
     
     
         19 . The non-transitory computer readable storage medium of  claim 15 , wherein the packet contains an extended transport header (ETH) indicating that the packet contains an atomic operation. 
     
     
         20 . The non-transitory computer readable storage medium of  claim 19 , wherein the ETH comprises a header that includes an atomic virtual address that specifies an address for an atomic operation and a base virtual address that specifies a first memory word of a buffer where data should be read from, or written to, at a destination node.

Cited by (0)

No later patents cite this yet.

References (0)

No backward citations on record.